Title: Critical Architecture Advice: Real-Time Ola/Uber-Style App Processes Dropping on Android Production (Capacitor JS)
Hi everyone,
I am a solo founder from India running a Government-recognized tech startup. I have invested my life savings (2 Lakh INR) into building a highly ambitious, location-based service marketplace. I am stuck at the absolute launch phase and desperately need verification on whether these are framework limits or code defects.
OUR CORE TECH WORKFLOW:
Our vision is an Ola, Uber, and Rapido-style instant-matching technology but applied to every major service category (with digital records for jobs).
1. Service workers can fix their availability slots or stay open for Instant Bookings.
2. For Instant Bookings: The request pings near-by workers instantly. The worker gets an interface to Accept, Reject, or send a Counter Offer to the customer.
THE CRITICAL PRODUCTION PROBLEMS (Post Google Play Store Release):
The app works fine in local development (Debug Mode). But now that it is LIVE on the Google Play Store (Release Build), it is failing on multiple budget devices (Xiaomi, Realme, Vivo, Oppo) with these bugs:
1. WORKER DISAPPEARANCE ON APP CLOSE:
When a worker toggles “Online” or fixes a slot, everything works. But the exact second the worker swipes up and kills the application from “Recent Apps” (Background app kill), they completely disappear from the customer map and database. Websockets/polling drop immediately. My developer says a real-time driver tracking/matching architecture cannot run stably on Capacitor JS. Is it impossible to run a persistent foreground matching service here?
2. STEP COMPLETION RESETTING TO “NOT COMPLETED”:
When a worker fixes a slot or completes a specific setup step, the UI temporarily reflects it. But upon navigating away or re-opening the view, it randomly reverts back to showing “Not Completed”.
3. SLUGGISH ONLINE/OFFLINE TOGGLE:
The real-time status toggle lags terribly, freezes, or fails to respond on live release builds, completely ruining the instant matching flow.
4. “FAILURE TO FETCH” NETWORK ERROR:
Production users are constantly hitting “Failure to Fetch” errors on their screens, which never occurred in local debug environments.
5. NATIVE CAMERA INTENT:
Triggering the camera frequently falls back to opening the system Photo Gallery instead of the physical lens interface.
I am trapped right before launch with my entire capital deployed. Please tell me plainly: Are these bugs actually caused by Capacitor JS limitations, or is it 100% possible to fix and run this kind of real-time Uber-style architecture smoothly on Capacitor?
Thank you for rescuing a drowning founder.